Pith. sign in

REVIEW 3 major objections 5 minor 47 references

Unlocking the Power of SAM 2 for Few-Shot Segmentation

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read FSSAM reuses SAM 2's video memory matching for few-shot segmentation by matching query features against pseudo query memories instead of support features, and reports state-of-the-art mIoU on PASCAL-5i and COCO-20i.

desk verdict A genuinely good SAM 2 adaptation for few-shot segmentation, but a ten-point fold-1 mIoU discrepancy across tables makes the headline 81.0 and the 4.2% margin unsupported until fixed. read the letter →

arxiv 2505.14100 v2 pith:HR2S2KVX submitted 2025-05-20 cs.CV

Qianxiong Xu , Lanyun Zhu , Xuanyi Liu , Guosheng Lin , Cheng Long , Ziyue Li , Rui Zhao This is my paper · ORCID
classification cs.CV
keywords queryfeaturesmemorysegmentationdesignmatchingclass-agnosticclasses
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Few-shot segmentation is the task of cutting out objects of a class the model has never been trained on, using only one or a few labelled examples. The hard part is that the model tends to overfit to the classes it saw during training. This paper starts from SAM 2, a video model that can track an object across frames. The authors observe that SAM 2 is good at matching the same object in two frames, but few-shot segmentation asks it to match two different objects of the same class, which is harder. Their solution is to first build rough 'pseudo' masks of the query image using features from another foundation model, DINOv2. Those rough masks are fed into SAM 2's memory encoder as if they were real video-frame masks, so SAM 2 matches the query image with itself rather than with the support image.

Because the rough masks are imperfect, two extra modules fix them. Iterative Memory Refinement gradually copies foreground features from a more complete but noisier mask into a cleaner mask. Support-Calibrated Memory Attention uses the clean support foreground as a reference to lower the attention paid to background pixels that leaked into the memory. On the standard PASCAL-5i and COCO-20i benchmarks, the final system reports 81.0% and 62.3% mean IoU in the 1-shot setting, ahead of previous methods. The authors also report ablations showing each module helps, and they include error bars and additional datasets in the appendix.

Extended reading notes

Core claim

The load-bearing assertion, from Section 5.2: 'FSSAM can surpass other baselines by large margins, setting new state-of-the-arts. For example, the 1-shot mIoU score is 4.2% better than that of the best baseline FounFSS, and the FB-IoU score can reach 89.4%.' If the paper is correct, adapting SAM 2 via pseudo query memories, iterative refinement, and support-calibrated attention yields 1-shot mIoU of 81.0 on PASCAL-5i and 62.3 on COCO-20i, with each designed module contributing to the gain in the ablation study.

Load-bearing premise

The design rests on the assumption that the pseudo query masks generated by DINOv2 prototypes are reliable enough to serve as SAM 2 mask prompts: if the prior completely misses the foreground or is dominated by background, the encoded memory is wrong and the method has no correction path. The authors acknowledge this in Appendix E ('the real FG is completely uncovered'). A second load-bearing premise is that cosine similarity between the support foreground prototype and query memory features cleanly separates query FG from query BG in IMR Eq. (8) and SCMA Eq. (12)-(13); the paper shows examples but gives no quantitative failure analysis.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes FSSAM, a method to adapt SAM 2 for few-shot segmentation. It identifies an incompatibility between SAM 2's same-object video matching and the different-object matching required by FSS, and addresses it with three components: a Pseudo Prompt Generator (PPG) that creates pseudo query masks encoded as query memory, an Iterative Memory Refinement (IMR) that progressively fuses foreground features into the discriminative memory, and a Support-Calibrated Memory Attention (SCMA) that suppresses background features in the memory during cross attention. The method is evaluated on PASCAL-5i, COCO-20i, LVIS-92i, and PASCAL-Part, reporting 1-shot mIoU of 81.0 on PASCAL-5i and 62.3 on COCO-20i, with component ablations, error bars, and qualitative visualizations. The authors also release code and state that the proposed modules introduce no additional learnable parameters.

Significance. If the reported results are correct, the paper makes a useful contribution by showing how a video foundation model (SAM 2) can be adapted to few-shot segmentation through a pseudo-prompt mechanism, rather than treating SAM 2 as a black-box promptable segmenter. The design is transparent, the modules are simple and parameter-free, and the appendices provide error bars, multi-episode testing, model-size studies, and additional datasets, which are all strengths. However, the central quantitative claim is currently undermined by an unreconciled inconsistency in the PASCAL-5i fold-1 score, and by test-set-based selection of the IMR iteration count. These issues must be resolved before the state-of-the-art claim can be trusted.

major comments (3)
  1. [Section 5.2, Table 1 vs. Table 3 and Table 11] The same final configuration (PPG+IMR+SCMA, fine-tuned) reports fold-1 mIoU of 84.9 in Table 1, but 74.9 in Table 3 (the full ablation row) and in Table 11 (FSSAM w/ FT). The four-fold mean in Tables 3 and 11 is still reported as 81.0, which is arithmetically impossible with the stated per-fold values (81.6+74.9+81.6+76.0)/4 = 78.5, not 81.0. This inconsistency directly affects the headline claim of a 4.2% margin over FounFSS: if the 74.9 value were correct, the margin would be about 1.7 points. The authors must reconcile these numbers, state which value is correct, and correct the affected means and conclusions. This is a load-bearing issue and cannot be left as an unexplained discrepancy.
  2. [Section 5.3, Table 4] The IMR iteration count n is selected as the peak of the test-set performance curve: Table 4 reports n=1,2,3,4 on PASCAL-5i and the paper sets n=3 because it gives the highest mean (81.0). This is test-set model selection: the hyperparameter is chosen using the same benchmark on which the final comparison is made, which can inflate the reported accuracy. The authors should either select n on a validation split, report the results for all values of n in the comparison with baselines, or clearly characterize the sensitivity. The effect is not negligible: n=1 gives 79.3 and n=2 gives 80.3, so the choice of n contributes up to 1.7 points to the headline number. Similarly, the SCMA scaling factor alpha in Eq. (13) is empirically set to 10; the selection procedure for alpha should be described.
  3. [Appendix C.5, Table 10] Table 10 is internally inconsistent. For Matcher, the four fold values 37.1, 32.4, 33.7, 38.1 average to 35.3, not the reported 42.9; for FSSAM, the values 34.7, 37.8, 37.2, 41.1 average to 37.7, not the reported 46.4. Moreover, the FSSAM fold values in Table 10 are identical to the first four folds of the LVIS-92i results in Table 9, which suggests that the rows may have been mis-copied or that the mean column corresponds to a different experimental setting. Since the paper uses these results to claim generalizability beyond PASCAL and COCO, the table must be corrected or the experimental protocol clarified.
minor comments (5)
  1. [Section 4.1] The Pseudo Prompt Generator is described as a contribution, but Equations (2)-(4) follow AENet's FG/BG prototype subtraction almost directly. The authors do cite AENet, but the novelty should be scoped more carefully: the contribution is the use of such priors as pseudo mask prompts for SAM 2, not the prior-generation mechanism itself.
  2. [Equations (8) and (13)] The notation AQQ is reused for different quantities: in Eq. (8) it is a similarity map between M_em^FG_Q and a prototype, while in Eq. (12)-(13) it is the query-memory attention map. Please use distinct symbols or clearly distinguish the two roles to improve readability.
  3. [Appendix C.1, Table 5] The COCO-20i fold-level error bars are fairly large (e.g., fold-0 ranges from 55.8 to 61.8 across seeds), yet the main Table 2 reports only the seed-321 result. Consider reporting the mean over multiple seeds in the main comparison to avoid overstating precision.
  4. [Table 3] The ablation table rows labeled only with checkmarks are hard to parse because the column order (PPG, IMR, SCMA) is clear in the header but the rows with two checkmarks do not indicate which one is omitted. Explicitly labeling each row, or adding a note, would improve clarity.
  5. [Abstract and Section 5.2] The phrase 'the 1-shot mIoU can be 4.2% better than the best baseline' is stated as a general result, but it refers specifically to PASCAL-5i. Please state the dataset explicitly in the abstract to avoid ambiguity.
Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The central empirical claim relies mainly on standard supervised training plus two domain assumptions: SAM 2's video matching is same-instance matching, and DINOv2 prototype priors are accurate enough to prompt SAM 2. The only hand-set numeric parameters that directly affect the headline results are the IMR iteration count n=3, chosen on test data, and the SCMA scaling factor alpha=10. No invented entities are introduced; the pseudo query memory is a method construct, not an independently evidenced object.

free parameters (2)
  • SCMA scaling factor alpha = 10
    Empirically set in Eq. (13); controls how strongly the support-calibrated BG suppression biases the attention scores, and no sensitivity analysis is provided.
  • IMR iteration count n = 3
    Chosen from Table 4 because n=3 gives the best PASCAL-5i mIoU; this is effectively a hyperparameter selected on the test folds.
assumptions (3)
  • domain assumption SAM 2's memory attention learned same-object FG-FG matching during video pretraining, and this knowledge is incompatible with matching two different object instances of the same class.
    Stated in Section 1 and Section 4.1 as the motivation for PPG; supported only by the ablation that naive SAM 2 achieves 69.4 versus 78.2 with PPG, not by direct analysis of SAM 2's internals.
  • domain assumption DINOv2 global-average-pooled support prototypes yield reliable FG and BG prior masks for novel classes.
    PPG in Section 4.1 copies the AENet prototype pipeline; the paper provides qualitative examples but no quantitative prior quality measure, and Appendix E admits complete failure cases.
  • domain assumption Cosine similarity between the support FG prototype and query memory features separates query FG from query BG well enough to suppress BG attention.
    Used in IMR Eq. (7)-(8) and SCMA Eq. (12)-(13); the paper gives visual examples but no quantitative failure analysis of when this separation breaks.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Unlocking the Power of SAM 2 for Few-Shot Segmentation." pith.science (2026). https://pith.science/paper/HR2S2KVX

@misc{pith2026250514100,
  author       = {Pith},
  title        = {Pith review of: Unlocking the Power of SAM 2 for Few-Shot Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HR2S2KVX}},
  note         = {Machine review of arXiv:2505.14100}
}
abstract

Few-Shot Segmentation (FSS) aims to learn class-agnostic segmentation on few classes to segment arbitrary classes, but at the risk of overfitting. To address this, some methods use the well-learned knowledge of foundation models (e.g., SAM) to simplify the learning process. Recently, SAM 2 has extended SAM by supporting video segmentation, whose class-agnostic matching ability is useful to FSS. A simple idea is to encode support foreground (FG) features as memory, with which query FG features are matched and fused. Unfortunately, the FG objects in different frames of SAM 2's video data are always the same identity, while those in FSS are different identities, i.e., the matching step is incompatible. Therefore, we design Pseudo Prompt Generator to encode pseudo query memory, matching with query features in a compatible way. However, the memories can never be as accurate as the real ones, i.e., they are likely to contain incomplete query FG, and some unexpected query background (BG) features, leading to wrong segmentation. Hence, we further design Iterative Memory Refinement to fuse more query FG features into the memory, and devise a Support-Calibrated Memory Attention to suppress the unexpected query BG features in memory. Extensive experiments have been conducted on PASCAL-5$^i$ and COCO-20$^i$ to validate the effectiveness of our design, e.g., the 1-shot mIoU can be 4.2% better than the best baseline.

Figures

Figures reproduced from arXiv: 2505.14100 by the authors.

Figure 1
Figure 1. Illustrations of (a) video data, (b) simple use of SAM 2, (c) our main idea, (d) prior masks and our Iterative Memory Refinement, and (e) our Support-Calibrated Memory Attention. In (a), SAM 2’s learned knowledge is same-objects matching. In (b), the objects in FSS are different, posing challenges to use SAM 2’s knowledge. In (c), we generate prior masks and encode them as pseudo query memories to enable same-object… view at source ↗
Figure 2
Figure 2. Overview of FSSAM, which includes: (1) Pseudo Prompt Generator generates a pair of FG (with more complete FG but more wrongly activated BG regions) and Disc priors (with less complete FG yet less wrongly activated BG regions) for encoding pseudo query memories MemF G Q and MemDisc Q ; (2) Iterative Memory Refinement aims to iteratively complement FG features from MemF G Q to MemDisc Q ; and (3) Support-Calibrated Me… view at source ↗
Figure 3
Figure 3. Details of Iterative Memory Refinement (IMR). IMR refines Disc memory MemDisc Q , by measuring its similarity with memory MemF G Q , and incorporating sufficient query FG features from the latter into the former. Meanwhile, support memory MemS (only FG) is used to prevent from fusing many BG features from MemF G S , so the refined Disc memory MemDisc Q can have more complete FG, while still include few BG features. … view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Illustrations of Support-Calibrated Memory Attention (SCMA). We only show cross attention in this figure. When per￾forming cross attention between FQ and MemDisc Q (FG&BG), the irrelevant memory (BG) will be suppressed by MemS (FG). Motivation. Remind that our goal is …
Figure 5
Figure 5. Figure 5: Qualitative illustrations of (a) query and support samples, (b) pseudo priors (mask prompts), (c) iterative memory refinement, and (d) outputs. We plot some rectangles to highlight some FG and BG areas. In (b), FG prior appears to have more complete FG but more wrongly…
Figure 6
Figure 6. Figure 6: Visual impacts of SCMA. The last two columns are the cross attention scores of the blue star. Orange rectangles represent some challenging areas that should be classified as BG. Visualizations of SCMA. An example is depicted in Fig￾ure 6, where the second column repres…
Figure 7
Figure 7. Figure 7: Overview of the simple idea to use SAM 2 for FSS. First of all, query and support features FQ and FS are extracted by the image encoder. Then, support features FS will be encoded as support memory MemS, taking support mask MS as mask prompt. After that, query features …
Figure 8
Figure 8. Figure 8: More visualizations of prior masks (i.e., pseudo mask prompts) and outputs. Support Query FG Prior Disc Prior Output [PITH_FULL_IMAGE:figures/full_fig_p016_8.png]
Figure 9
Figure 9. Figure 9: More visualizations of prior masks (i.e., pseudo mask prompts) and outputs. 16 [PITH_FULL_IMAGE:figures/full_fig_p016_9.png]
Figure 10
Figure 10. Figure 10: More visualizations of Iterative Memory Refinement. Support Query FG Prior Disc Prior 𝑛𝑛 = 1 𝑛𝑛 = 2 𝑛𝑛 = 3 𝑛𝑛 = 4 (a) Inputs and Labels (b) Pseudo Mask Prompts (c) Iterative Memory Refinement [PITH_FULL_IMAGE:figures/full_fig_p017_10.png]
Figure 11
Figure 11. Figure 11: More visualizations of Iterative Memory Refinement. 17 [PITH_FULL_IMAGE:figures/full_fig_p017_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 36 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    Relevant intrinsic feature enhancement network for few-shot semantic segmentation

    Bao, X., Qin, J., Sun, S., Wang, X., and Zheng, Y. Relevant intrinsic feature enhancement network for few-shot semantic segmentation. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pp.\ 765--773, 2024

    work page 2024

  3. [3]

    High-Performance Few-Shot Segmentation with Foundation Models: An Empirical Study

    Chang, S., Zhang, L., and Lu, H. High-performance few-shot segmentation with foundation models: An empirical study. arXiv preprint arXiv:2409.06305, 2024

    work page Pith review arXiv 2024

  4. [4]

    No Re-Train, More Gain: Upgrading Backbones with Diffusion model for Pixel-Wise and Weakly-Supervised Few-Shot Segmentation

    Chen, S., Meng, F., Wu, C., Wei, H., Zhang, R., Wu, Q., Xu, L., and Li, H. No re-train, more gain: Upgrading backbones with diffusion model for few-shot segmentation. arXiv preprint arXiv:2407.16182, 2024

    work page Pith review arXiv 2024

  5. [5]

    Detect what you can: Detecting and representing objects using holistic models and body parts

    Chen, X., Mottaghi, R., Liu, X., Fidler, S., Urtasun, R., and Yuille, A. Detect what you can: Detecting and representing objects using holistic models and body parts. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 1971--1978, 2014

    work page 1971

  6. [6]

    K., Winn, J., and Zisserman, A

    Everingham, M., Van Gool, L., Williams, C. K., Winn, J., and Zisserman, A. The pascal visual object classes (voc) challenge. International journal of computer vision, 88 0 (2): 0 303--338, 2010

    work page 2010

  7. [7]

    Self-support few-shot semantic segmentation

    Fan, Q., Pei, W., Tai, Y.-W., and Tang, C.-K. Self-support few-shot semantic segmentation. In Computer Vision--ECCV 2022: 17th European Conference, Tel Aviv, Israel, October 23--27, 2022, Proceedings, Part XIX, pp.\ 701--719. Springer, 2022

    work page 2022

  8. [8]

    and Dao, T

    Gu, A. and Dao, T. Mamba: Linear-time sequence modeling with selective state spaces. arXiv preprint arXiv:2312.00752, 2023

    arXiv 2023

Show all 47 references
  1. [9]

    Lvis: A dataset for large vocabulary instance segmentation

    Gupta, A., Dollar, P., and Girshick, R. Lvis: A dataset for large vocabulary instance segmentation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 5356--5364, 2019

    2019

  2. [10]

    Simultaneous detection and segmentation

    Hariharan, B., Arbel \'a ez, P., Girshick, R., and Malik, J. Simultaneous detection and segmentation. In European conference on computer vision, pp.\ 297--312. Springer, 2014

    2014

  3. [11]

    Cost aggregation with 4d convolutional swin transformer for few-shot segmentation

    Hong, S., Cho, S., Nam, J., Lin, S., and Kim, S. Cost aggregation with 4d convolutional swin transformer for few-shot segmentation. In European Conference on Computer Vision, pp.\ 108--126. Springer, 2022

    2022

  4. [12]

    Hu, T., Yang, P., Zhang, C., Yu, G., Mu, Y., and Snoek, C. G. Attention-based multi-context guiding for few-shot semantic segmentation. In Proceedings of the AAAI conference on artificial intelligence, volume 33, pp.\ 8441--8448, 2019

    2019

  5. [13]

    Msanet: Multi-similarity and attention guidance for boosting few-shot segmentation

    Iqbal, E., Safarov, S., and Bang, S. Msanet: Multi-similarity and attention guidance for boosting few-shot segmentation. arXiv preprint arXiv:2206.09667, 2022

    2022 arXiv

  6. [14]

    C., Lo, W.-Y., et al

    Kirillov, A., Mintun, E., Ravi, N., Mao, H., Rolland, C., Gustafson, L., Xiao, T., Whitehead, S., Berg, A. C., Lo, W.-Y., et al. Segment anything. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 4015--4026, 2023

    2023

  7. [15]

    Lin, T.-Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Doll \'a r, P., and Zitnick, C. L. Microsoft coco: Common objects in context. In European conference on computer vision, pp.\ 740--755. Springer, 2014

    2014

  8. [16]

    Fecanet: Boosting few-shot semantic segmentation with feature-enhanced context-aware network

    Liu, H., Peng, P., Chen, T., Wang, Q., Yao, Y., and Hua, X.-S. Fecanet: Boosting few-shot semantic segmentation with feature-enhanced context-aware network. IEEE Transactions on Multimedia, 2023 a

    2023

  9. [17]

    Matcher: Segment anything with one shot using all-purpose feature matching

    Liu, Y., Zhu, M., Li, H., Chen, H., Wang, X., and Shen, C. Matcher: Segment anything with one shot using all-purpose feature matching. arXiv preprint arXiv:2305.13310, 2023 b

    2023 arXiv

  10. [18]

    A simple image segmentation framework via in-context examples

    Liu, Y., Jing, C., Li, H., Zhu, M., Chen, H., Wang, X., and Shen, C. A simple image segmentation framework via in-context examples. arXiv preprint arXiv:2410.04842, 2024

    2024 arXiv

  11. [19]

    Fully convolutional networks for semantic segmentation

    Long, J., Shelhamer, E., and Darrell, T. Fully convolutional networks for semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 3431--3440, 2015

    2015

  12. [20]

    V-net: Fully convolutional neural networks for volumetric medical image segmentation

    Milletari, F., Navab, N., and Ahmadi, S.-A. V-net: Fully convolutional neural networks for volumetric medical image segmentation. In 2016 fourth international conference on 3D vision (3DV), pp.\ 565--571. IEEE, 2016

    2016

  13. [21]

    Attention-based joint detection of object and semantic part

    Morabia, K., Arora, J., and Vijaykumar, T. Attention-based joint detection of object and semantic part. arXiv preprint arXiv:2007.02419, 2020

    2007 arXiv

  14. [22]

    and Todorovic, S

    Nguyen, K. and Todorovic, S. Feature weighting and boosting for few-shot segmentation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 622--631, 2019

    2019

  15. [23]

    Dinov2: Learning robust visual features without supervision

    Oquab, M., Darcet, T., Moutakanni, T., Vo, H., Szafraniec, M., Khalidov, V., Fernandez, P., Haziza, D., Massa, F., El-Nouby, A., et al. Dinov2: Learning robust visual features without supervision. arXiv preprint arXiv:2304.07193, 2023

    2023 arXiv

  16. [24]

    S., and Heo, J.-P

    Park, S., Lee, S., Hyun, S., Seong, H. S., and Heo, J.-P. Task-disruptive background suppression for few-shot segmentation. In Proceedings of the AAAI conference on artificial intelligence, volume 38, pp.\ 4442--4449, 2024

    2024

  17. [25]

    Hierarchical dense correlation distillation for few-shot segmentation

    Peng, B., Tian, Z., Wu, X., Wang, C., Liu, S., Su, J., and Jia, J. Hierarchical dense correlation distillation for few-shot segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 23641--23651, 2023

    2023

  18. [26]

    Sam 2: Segment anything in images and videos

    Ravi, N., Gabeur, V., Hu, Y.-T., Hu, R., Ryali, C., Ma, T., Khedr, H., R \"a dle, R., Rolland, C., Gustafson, L., et al. Sam 2: Segment anything in images and videos. arXiv preprint arXiv:2408.00714, 2024

    2024 arXiv

  19. [27]

    One-shot learning for semantic segmentation

    Shaban, A., Bansal, S., Liu, Z., Essa, I., and Boots, B. One-shot learning for semantic segmentation. arXiv preprint arXiv:1709.03410, 2017

    2017 arXiv

  20. [28]

    Dense cross-query-and-support attention weighted mask aggregation for few-shot segmentation

    Shi, X., Wei, D., Zhang, Y., Lu, D., Ning, M., Chen, J., Ma, K., and Zheng, Y. Dense cross-query-and-support attention weighted mask aggregation for few-shot segmentation. In European Conference on Computer Vision, pp.\ 151--168. Springer, 2022

    2022

  21. [29]

    Vrp-sam: Sam with visual reference prompt

    Sun, Y., Chen, J., Zhang, S., Zhang, X., Chen, Q., Zhang, G., Ding, E., Wang, J., and Li, Z. Vrp-sam: Sam with visual reference prompt. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 23565--23574, 2024

    2024

  22. [30]

    Prior guided feature enrichment network for few-shot segmentation

    Tian, Z., Zhao, H., Shu, M., Yang, Z., Li, R., and Jia, J. Prior guided feature enrichment network for few-shot segmentation. IEEE transactions on pattern analysis and machine intelligence, 2020

    2020

  23. [31]

    N., Kaiser, ., and Polosukhin, I

    Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, ., and Polosukhin, I. Attention is all you need. Advances in neural information processing systems, 30, 2017

    2017

  24. [32]

    Few-shot semantic segmentation with democratic attention networks

    Wang, H., Zhang, X., Hu, Y., Yang, Y., Cao, X., and Zhen, X. Few-shot semantic segmentation with democratic attention networks. In European Conference on Computer Vision, pp.\ 730--746. Springer, 2020

    2020

  25. [33]

    Adaptive fss: a novel few-shot segmentation framework via prototype enhancement

    Wang, J., Li, J., Chen, C., Zhang, Y., Shen, H., and Zhang, T. Adaptive fss: a novel few-shot segmentation framework via prototype enhancement. In Proceedings of the AAAI conference on artificial intelligence, volume 38, pp.\ 5463--5471, 2024 a

    2024

  26. [34]

    Rethinking prior information generation with clip for few-shot segmentation

    Wang, J., Zhang, B., Pang, J., Chen, H., and Liu, W. Rethinking prior information generation with clip for few-shot segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 3941--3951, 2024 b

    2024

  27. [35]

    H., Zou, Y., Zhou, D., and Feng, J

    Wang, K., Liew, J. H., Zou, Y., Zhou, D., and Feng, J. Panet: Few-shot image semantic segmentation with prototype alignment. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 9197--9206, 2019

    2019

  28. [36]

    Focus on query: Adversarial mining transformer for few-shot segmentation

    Wang, Y., Luo, N., and Zhang, T. Focus on query: Adversarial mining transformer for few-shot segmentation. Advances in Neural Information Processing Systems, 36: 0 31524--31542, 2023 a

    2023

  29. [37]

    Rethinking the correlation in few-shot segmentation: A buoys view

    Wang, Y., Sun, R., and Zhang, T. Rethinking the correlation in few-shot segmentation: A buoys view. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 7183--7192, 2023 b

    2023

  30. [38]

    Xiong, Z., Li, H., and Zhu, X. X. Doubly deformable aggregation of covariance matrices for few-shot segmentation. In Computer Vision--ECCV 2022: 17th European Conference, Tel Aviv, Israel, October 23--27, 2022, Proceedings, Part XX, pp.\ 133--150. Springer, 2022

    2022

  31. [39]

    Self-calibrated cross attention network for few-shot segmentation

    Xu, Q., Zhao, W., Lin, G., and Long, C. Self-calibrated cross attention network for few-shot segmentation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 655--665, 2023

    2023

  32. [40]

    C., Long, C., Li, Z., and Zhao, R

    Xu, Q., Lin, G., Loy, C. C., Long, C., Li, Z., and Zhao, R. Eliminating feature ambiguity for few-shot segmentation. In European Conference on Computer Vision, pp.\ 416--433. Springer, 2024 a

    2024

  33. [41]

    Hybrid mamba for few-shot segmentation

    Xu, Q., Liu, X., Zhu, L., Lin, G., Long, C., Li, Z., and Zhao, R. Hybrid mamba for few-shot segmentation. Advances in Neural Information Processing Systems, 37: 0 73858--73883, 2024 b

    2024

  34. [42]

    Mianet: Aggregating unbiased instance and general information for few-shot semantic segmentation

    Yang, Y., Chen, Q., Feng, Y., and Huang, T. Mianet: Aggregating unbiased instance and general information for few-shot semantic segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 7131--7140, 2023

    2023

  35. [43]

    Bridge the points: Graph-based few-shot segment anything semantically

    Zhang, A., Gao, G., Jiao, J., Liu, C., and Wei, Y. Bridge the points: Graph-based few-shot segment anything semantically. Advances in Neural Information Processing Systems, 37: 0 33232--33261, 2024

    2024

  36. [44]

    Pyramid graph networks with connection attentions for region-based one-shot semantic segmentation

    Zhang, C., Lin, G., Liu, F., Guo, J., Wu, Q., and Yao, R. Pyramid graph networks with connection attentions for region-based one-shot semantic segmentation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 9587--9595, 2019 a

    2019

  37. [45]

    Canet: Class-agnostic segmentation networks with iterative refinement and attentive few-shot learning

    Zhang, C., Lin, G., Liu, F., Yao, R., and Shen, C. Canet: Class-agnostic segmentation networks with iterative refinement and attentive few-shot learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 5217--5226, 2019 b

    2019

  38. [46]

    Few-shot segmentation via cycle-consistent transformer

    Zhang, G., Kang, G., Yang, Y., and Wei, Y. Few-shot segmentation via cycle-consistent transformer. Advances in Neural Information Processing Systems, 34: 0 21984--21996, 2021

    2021

  39. [47]

    Llafs: When large language models meet few-shot segmentation

    Zhu, L., Chen, T., Ji, D., Ye, J., and Liu, J. Llafs: When large language models meet few-shot segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 3065--3075, 2024

    2024

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.